JP5296765B2 - JPEG compression device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To solve a problem in which: a JPEG data size does not easily exceed a transmission band even when video images suddenly vary and an image data size increases. <P>SOLUTION: A JPEG compression device includes: a moving average calculation part 42 for calculating the moving average of M pieces of continuous frames of the past; a moving object identification part 43 for comparing the JPEG data sizes of the (N-1)th frame and Nth frame of moving images and determining that a moving object is present when the extreme reduction of the JPEG data size occurs; a compressibility setting part 45 for setting compressibility on the basis of the JPEG data size of the (N-1)th frame and information of the moving average calculation part 42 and the moving object identification part 43; a quantization table generation part 46 for generating a quantization table by the set compressibility; and a JPEG encoding part 41 for executing JPEG compression of the (N+1)th frame of the frames on the basis of the generated quantization table. When the moving object identification part 43 determines that the moving object is present, the compressibility determination part 45 sets the compressibility at the maximum value. <P>COPYRIGHT: (C)2012,JPO&amp;INPIT

Description

  The present invention relates to a JPEG compression apparatus that compresses JPEG data, and more particularly to a JPEG compression apparatus that can transmit a moving image compressed according to the JPEG standard at a predetermined frame rate so as to be within a constant bit rate.

  JPEG (Joint Picture Expert Group) is a standard method for compressing and expanding color still images, and the compression rate of the current image frame is determined from the data size of the previous image frame in multiple consecutive image frames that make up a moving image. There is known a JPEG compression apparatus configured so that the bit rate of a compressed moving image falls within a certain value (see, for example, Patent Document 1).

Japanese Patent Laid-Open No. 2007-60498

  The JPEG compression device described in Patent Document 1 controls the amount of data to be transmitted based on the premise that the data size of the image of the immediately preceding image frame and that of the current image frame hardly change. For this reason, when the JPEG data size has decreased in the immediately preceding image frame, control is performed to increase the data amount by switching the quantization table to a high-quality image in the next image frame.

  However, in the case of a frame rate of about 30 frames / second, which is often used for intercom entrance cameras and surveillance cameras, the video changes drastically when a visitor or suspicious person enters or leaves the camera. As a result, the JPEG data size changes greatly.

This situation will be specifically described with reference to the drawings. FIG. 4 shows the transition of the video state when the moving object stops on the screen, and FIG. 4A shows an image before the moving object enters the screen. FIG. 4B shows an image in which the moving object is moving near the center of the screen, and FIG. 4C shows an image in which the moving object is stopped in the imaging screen. FIG. 6 shows the transition of the JPEG data size in the transition of FIG.
In this way, when a situation occurs in which an object such as a person enters the camera from the side, moves and stops near the front of the camera, before the moving object enters the screen, it is indicated by M1 in FIG. Although the data size is such that the image is blurred and monotonous while moving due to the frame-in of the moving object, the data size is significantly reduced to M2 in FIG. Therefore, the quantization table is switched to high image quality in order to control to increase the data size.

After that, when the operation is stopped, there is no blurring of the image and the image becomes complicated, so that the data size is greatly increased to M3 in FIG. At this time, since the quantization table is switched to a high image quality, the data size greatly exceeds the transmission band.
In this way, when the moving object stops in the frame, the video is thinned out, and a state where the moving image is stopped for a moment is displayed on the monitor.

FIG. 5 shows the transition of the video state when a moving object crosses the screen in the front of a relatively complex background, and FIG. 5A shows an image immediately after the moving object enters the screen. FIG. 5B shows an image in which the moving object is moving on the screen, and FIG. 5C shows an image immediately after the moving object has left the screen without stopping. FIG. 7 shows the transition of the JPEG data size in the transition of FIG.
Even when a moving object such as a person crosses the frame in front of the camera in this way, the JPEG data size indicated by M4 in FIG. Then, as indicated by M5 in FIG. 7, the JPEG data size is greatly reduced. After that, when all the objects are removed from the screen, the original complicated background is displayed, so that the JPEG data size increases at a stretch as indicated by M6 in FIG. As a result, as in the case where the moving object stops, a situation occurs in which the JPEG data size exceeds the transmission band.
In both states, the frequency of occurrence is high, and when trying to solve this problem using the conventional method, the entire quantization table has to be shifted in a direction in which the image quality decreases.

  In the conventional method, since the determination is based on the JPEG data size of the entire image frame, even if a partial increase in the JPEG data size due to a part of the video fluctuates, the JPEG data size of the entire frame is not affected. If there was little fluctuation, the quantization table was not switched. However, even if the JPEG data size of the entire frame does not exceed the transmission band, it often occurs that the transmission band is instantaneously exceeded when compressing complex parts of the video, so that the video is disturbed. It was. In order to cope with this, a frame buffer for storing JPEG data in units of frames had to be provided on the JPEG compression side.

  Therefore, in view of such problems, the present invention makes it difficult for the JPEG data size to exceed the transmission band even if the image data fluctuates and the image data size suddenly increases, and the frequency of instantaneous stop due to image thinning is reduced. JPEG compression apparatus that can be used, and by determining fluctuations in JPEG data size in units of blocks obtained by dividing an image frame, even if there is no frame buffer on the JPEG compression side, instantaneous JPEG data An object of the present invention is to provide a JPEG compression apparatus capable of preventing the transmission band from being exceeded due to an increase.

In order to solve the above-mentioned problems, the invention of claim 1 is a JPEG compression apparatus that converts video data into moving image data composed of JPEG data and outputs the video data, and includes M consecutive M (M is an integer) in the past. A moving average calculating unit that calculates a moving average of JPEG data sizes in an image frame, and comparing the JPEG data sizes of the N−1th (N is an integer) image frame and the Nth image frame of a moving image, Information on a moving object identification unit that determines that a moving object exists when a decrease in size exceeding a predetermined change amount occurs, and information on the JPEG data size, moving average calculation unit, and moving object identification unit of the (N−1) -th image frame Based on the compression rate setting unit for setting the compression rate of the (N + 1) th image frame, and the amount used for JPEG compression of the (N + 1) th image frame from the set compression rate A quantization table generating unit that generates a quantization table, a JPEG encoding unit that performs JPEG compression of an image frame and JPEG encodes video data based on the quantization table generated by the quantization table generating unit; A data transmission unit that deletes the image frame and sends a thinned frame when the data size of the JPEG data output by the JPEG encoding unit exceeds a predetermined size, and the compression rate setting unit is configured to identify the moving object If the unit determines that there is a moving object, the compression rate is set to a set maximum value, and the JPEG data output by the JPEG encoding unit is shifted to a low image quality.
This invention uses the fact that JPEG data is greatly reduced due to image blurring when the movement of an object is imaged, and the JPEG data size is reduced when the object is moved. However, in the present invention, on the contrary, the table is switched to the table with the lowest image quality, and the JPEG data size is shifted so that the JPEG data size does not exceed the transmission band even if the object stops later, and the preparation operation is performed. To do.
As a result, the image quality temporarily decreases, but the details of the image cannot be identified much when the object is moving and the blurred image is displayed, so there is almost no effect of the image quality deterioration, and the image changes rapidly. Even if the data size of the image frame increases, the JPEG data size does not easily exceed the transmission band, and the frequency of instantaneous stop due to video thinning decreases. Then, during normal times without large fluctuations in the data size, JPEG data having an image quality suitable for the data size is encoded and transmitted.

According to a second aspect of the present invention, in the configuration of the first aspect, the JPEG encoding unit sequentially encodes the input image frame in units of a predetermined number of pixels, and JPEG data for each encoded block. When the block determination unit has a block determination unit that compares the JPEG data size with a predetermined value and outputs the block while the JPEG data size exceeds the predetermined value, the block determination unit sends a compression rate to the compression rate setting unit. In this case, the data compression in the JPEG encoding unit is increased by one step to shift to low image quality.
According to this configuration, the JPEG data size is counted for each block to be encoded, and the quantization table switching determination is performed for each block. Therefore, even if there is almost no change in the JPEG data size in the entire image frame, if there is a sudden increase in JPEG data in a part of the image frame, it is detected and the quantization table is quickly switched to the low image quality side. As a result, it is possible to prevent the JPEG data from instantaneously exceeding the transmission band, and it is possible to transmit an image satisfactorily without having a frame buffer on the JPEG compression side.

According to the present invention, it is difficult to exceed the transmission band even if the video changes rapidly and the JPEG data size increases, and the frequency of instantaneous stop due to video thinning can be reduced. At the same time, deterioration of image quality due to JPEG compression can be minimized.
In addition, by judging even the block unit that divides the image frame into multiple, even if there is no large fluctuation in the entire JPEG data size, the situation where the instantaneous JPEG data exceeds the capacity of the transmission band and the video is disturbed at the data destination. The video can be transmitted satisfactorily without having a frame buffer on the JPEG compression side.

It is a block diagram of the intercom apparatus provided with the JPEG compression apparatus. It is a block diagram which shows an example of the JPEG compression apparatus which concerns on this invention. It is a flowchart which shows the flow of video data. It is an image state transition diagram by a moving object. It is an image state transition diagram by a moving object. FIG. 5 is a transition diagram of JPEG data size in the video state transition of FIG. 4. FIG. 6 is a transition diagram of JPEG data size in the video state transition of FIG. 5. It is an image data configuration of one frame.

DESCRIPTION OF EXEMPLARY EMBODIMENTS Hereinafter, embodiments of the invention will be described in detail with reference to the drawings. FIG. 1 shows an intercom apparatus, and FIG. 2 shows a block diagram of the main part. As an example, the JPEG compression apparatus according to the present invention is incorporated in an entrance unit of an intercom apparatus.
In FIG. 1, reference numeral 1 denotes an entrance slave set installed at the entrance, and reference numeral 2 denotes a living room set installed in the living room, and both are connected by a transmission line L <b> 1. The front door 1 includes a call button 11 for a visitor to call a resident, a camera 12 for capturing an image of the visitor, a microphone 13 and a speaker 14 for calling. The room master unit 2 includes a call button 21 for responding to a call, a microphone 22 and a speaker 23 for making a call, and a monitor 24 for displaying an image.

FIG. 2 shows blocks related to video data processing until the video captured by the camera 12 is displayed on the monitor 24. The block diagram of the front door 1 is a block diagram of the JPEG compression device 4. As shown in FIG. 2, the JPEG compression device 4 includes a JPEG encoding unit 41 that encodes (compresses) video data that is captured and output by the camera 12, a moving average calculation unit 42 that calculates a moving average, The data size is monitored in a moving object identification unit 43 that identifies a moving object based on a count value (data size) of JPEG compressed data and individual image blocks that form an image frame (hereinafter simply referred to as “frame”). A block determination unit 44; a compression rate determination unit 45 that determines a compression rate in the JPEG encoding unit 41; a quantization table generation unit 46 that generates a quantization table according to the determined compression rate; 2 is provided with a data transmission unit 47 for transmitting video data to 2.
The living room master unit 2 also includes a data transmission unit 50 that receives the transmitted video data and a JPEG decoding unit that decodes the JPEG data, and the video is displayed on the monitor 24 by the decoded moving image data.

The JPEG encoding unit 41 uses the quantization table sent from the quantization table generation unit 46 to encode and output moving image data output from the camera 12. Specifically, the video data output from the camera 12 is moving image data composed of continuous image data, and is divided into a large number of image blocks and subjected to JPEG encoding as shown in FIG.
FIG. 8 shows the structure of image data for one frame. As shown in FIG. 8, the frame is divided into blocks of MCU (for example, 8 × 8 pixels) and encoded.
Then, the count value of the JPEG data is sequentially output in units of MCU, and when the JPEG encoding of one frame is completed, the frame data that is the compressed video data is output to the data transmission unit 47.

  The moving average calculation unit 42 obtains the count value from the JPEG encoding unit 41, calculates the average value of the JPEG data size of the frame every predetermined M (for example, 4), and determines the compression rate as the result. Send to part 45.

  The moving object identification unit 43 obtains the count value from the JPEG encoding unit 41 and compares the JPEG data sizes of the immediately previous (N−1) th frame and the current (Nth) frame. As a result of the comparison, if the data size has decreased by a certain value or more, it is determined that some moving object is captured in the video, and a signal to that effect is output to the compression rate determination unit 45. This is performed to cope with a sudden increase in JPEG data size that occurs after the moving object as shown in FIGS. 5 and 6 is imaged.

  The block determination unit 44 receives the count value in units of blocks output when the JPEG encoding unit 41 performs JPEG encoding, and determines the data size. Specifically, the JPEG data size is compared with a predetermined threshold value from the obtained count value, and if it is equal to or larger than the set threshold value, the compression rate determination unit 45 waits for the current JPEG compression process to end. An instruction is sent to the JPEG encoding unit 41 to send a quantization table for reducing the one-step image quality. The purpose of this is to prevent the case where the video is locally complicated in the frame and instantaneously exceeds the transmission band even if the JPEG data size of the entire frame is within the transmission band and there is no problem in transmission. .

The compression rate determination unit 45 receives the count value of the (N−1) th frame immediately before being output by the JPEG encoding unit 41, and generates a quantization table in which the image quality is reduced by one step if the count value is equal to or greater than the threshold value. A signal is output to the conversion table 46. Also, the average value calculated by the moving average calculation unit 42 is compared with a predetermined threshold value, and if it is equal to or greater than the threshold value, a quantization table with a reduced one-step image quality is sent to the JPEG encoding unit 41. The moving average calculation unit 42 does not output a signal for changing the image quality when the calculated average value is less than the threshold value.
When the moving object identification unit 43 identifies the presence of a moving object, it issues a command to the quantization table generation unit 46 to generate a quantization table with the lowest image quality (maximum compression rate).
Further, in response to a signal output from the block determination unit 44, a signal for lowering the image quality by one level is output to the quantization table generation unit 46.

  If the data size of the frame transmitted from the JPEG encoding unit 41 is equal to or less than a preset threshold, the data transmission unit 47 transmits the frame to the room master unit 2 via the transmission line L1. If the threshold is exceeded, the frame data is discarded.

  The JPEG compression apparatus 4 configured in this way processes video data as follows. FIG. 3 shows a flow from when the camera takes an image to transmission to the room master unit 2, and will be described based on FIG. Video data (moving image data) captured by the camera 12 is output to the JPEG encoding unit 41 (S1). The JPEG encoding unit 41 to which the video data is input receives the quantization table data having the lowest image quality determined by the initial setting from the quantization table generation unit 46 in the first compression, and performs JPEG compression processing (S2). ).

The compression is performed in units of blocks, and the count value of JPEG data, which is the result of compression, is output each time (S3). The count value of the JPEG data is simultaneously sent to the moving average calculation unit 42, the moving object identification unit 43, the block determination unit 44, and the compression rate determination unit 45. In the second and subsequent compression processing, the JPEG compression processing is performed using the quantization table generated based on the previous JPEG data from the quantization table generation unit 46.
The JPEG data obtained by compressing the image data of one frame in this way is output to the data transmission unit 47 (S4).

  The JPEG data output to the data transmission unit 47 is determined by the data transmission unit 47 whether the data amount of the image frame is equal to or less than a predetermined threshold (S5). (S6). If it exceeds, the frame data is discarded (S7). The room master 2 receives the JPEG data at the data transmission unit 50, is decoded into video data at the JPEG decoding unit 51, and is displayed on the monitor 24.

On the other hand, when the obtained count value reaches the determined number of frames, the moving average calculation unit 42 calculates the average value of the JPEG data sizes of the frames so far, and sends the result to the compression rate determination unit 45. The compression rate determination unit 45 compares the predetermined threshold value with the average value.
As a result of this comparison, a signal is output to the quantization table generation unit 46 so as to generate a quantization table that increases the image quality by one level if it is equal to or less than the threshold value. In this way, compression is performed with the quantization table from the next (N + 1) th frame. Note that once the quantization table is switched, no determination is made in the direction of increasing the image quality until a moving average of the determined number of frames is calculated again. By this process, the quantization table can be optimized in accordance with the current image quality, and by first compressing with the lowest image quality, the JPEG size that exceeds the transmission band by the first compression is prevented. If the threshold value is exceeded, the quantization table change instruction is not issued.

  When the compression rate determination unit 45 receives the count value from the JPEG encoding unit 41, the compression rate determination unit 45 issues an instruction to lower the image quality by one step if the data size is larger than a predetermined value. Specifically, the JPEG data size of the immediately preceding (N−1) th frame is compared with a threshold value, and if the immediately preceding data size is greater than or equal to the threshold value, the quantization table is generated so as to reduce the one-step image quality. A signal is output to the generation unit 46. The JPEG encoding unit 41 uses the quantization table received from the quantization table generating unit 46 to perform compression using the quantization table from the next (N + 1) th frame. If the previous data size is less than the threshold value, the quantization table is not changed. This operation in the direction of decreasing the JPEG data size (decreasing the image quality) is performed by determining each time one frame is compressed.

The moving object identification unit 43 compares the JPEG data sizes of the immediately preceding (N−1) th frame and the current (Nth) frame based on the count value transmitted from the JPEG encoding unit 41, and exceeds a certain level. If there is a large decrease in size, it is determined that there is some moving object in the video, and if it is determined that there is a moving object, the compression ratio determination unit 45 is notified.
Receiving this notification, the compression rate determination unit 45 causes the quantization table generation unit 46 to generate a quantization table with the lowest image quality among the set quantization tables. The JPEG encoding unit 41 uses the quantization table generated from the quantization table generating unit 46 and encodes the next (N + 1) th frame with the quantization table.

The block determination unit 44 determines the data size when receiving the count value in the (N−1) th frame immediately before the output from the JPEG encoding unit 41. If there is a block exceeding the threshold, the image quality is lowered. An instruction is issued to the compression rate determination unit 45. As described above, the data size is determined in MCU units, and if it is equal to or greater than the threshold value, the compression rate determination unit 45 waits for the Nth (current) JPEG compression process to end and waits for the next (N + 1) th frame. To instruct the JPEG encoding unit 41 to send a quantization table for reducing the one-step image quality.
In this way, the JPEG encoding unit 41 performs encoding processing for the next N + 1th frame.

In this way, when a moving object is imaged, the quantization table is switched to the one with the lowest image quality, and the JPEG data size does not exceed the transmission bandwidth even if the object stops later and the image becomes complicated. Therefore, even if the video changes suddenly and the data size of the frame increases, the JPEG data size does not easily exceed the transmission band, and the frequency of instantaneous stop due to video thinning can be reduced. In addition, changing the quantization table to the one with the lowest image quality temporarily decreases the image quality, but the details of the image cannot be identified much when the object is moving and the blurred image is displayed. There is almost no impact of the decline.
Also, since the JPEG data size is counted for each block to be encoded and the quantization table switching determination is made for each block, even if there is almost no change in the JPEG data size in the entire image frame, the JPEG data is partially recorded in the image frame. The case where there is a sudden increase in data is detected and the quantization table is quickly switched to the low image quality side. As a result, it is possible to prevent the JPEG data from instantaneously exceeding the transmission band, and it is possible to transmit an image satisfactorily without having a frame buffer on the JPEG encoding side.

  In the above embodiment, the block is configured in units of 8 × 8 pixels. However, the size of each block configuring the frame is not limited to this number of pixels. For example, the number of pixels is a multiple of this block. It may consist of numbers.

  In the above embodiment, the configuration applied to the interphone device has been described. However, the JPEG compression device 4 of the present invention transmits a captured image of a camera via a communication path or a transmission path and displays it on a monitor or the like. In this case, it is preferable to display the video without stopping.

  1 .. Entrance child machine 2.. Living room master 4.. JPEG compression device 12.. Camera 41. JPEG encoding unit 42 42 Moving average calculation unit 43 Moving object identification unit 44... Block determination unit, 45... Compression rate determination unit, 46 .. Quantization table generation unit, 47.

Claims (2)

A JPEG compression device that converts video data into moving image data composed of JPEG data and outputs the video data,
A moving average calculating unit for calculating a moving average of JPEG data sizes in past M image frames (M is an integer);
When the JPEG data size of the N-1th image frame (N is an integer) and the Nth image frame of the moving image are compared, and the JPEG data size exceeds a predetermined change amount, a moving object is present. A moving object identification unit to determine;
A compression rate setting unit that sets a compression rate of the (N + 1) th image frame based on the information of the JPEG data size of the (N-1) th image frame, the moving average calculation unit, and the moving object identification unit;
A quantization table generating unit that generates a quantization table used for JPEG compression of the (N + 1) th image frame from the set compression rate;
A JPEG encoding unit that performs JPEG compression of an image frame and JPEG encodes the video data based on the quantization table generated by the quantization table generation unit;
A data transmission unit that deletes the image frame and sends a thinned frame when the data size of the JPEG data output by the JPEG encoding unit exceeds a predetermined size, and
When the moving object identification unit determines that a moving object exists, the compression rate setting unit sets the compression rate to a set maximum value, and shifts the JPEG data output from the JPEG encoding unit to low image quality. JPEG compression apparatus characterized by this. The JPEG encoding unit sequentially encodes the input image frame in units of a predetermined number of pixels, and counts and outputs the JPEG data size for each encoded block,
A block determination unit that compares the JPEG data size with a predetermined value;
When there is a block whose JPEG data size exceeds a predetermined value, the block determination unit issues an instruction to increase the compression rate to the compression rate setting unit, and the data compression in the JPEG encoding unit is increased by one step to shift to low image quality. The JPEG compression apparatus according to claim 1, wherein:

Images